VistA suggests substance codes based on user input, but the user selects the appropriate code. Each suggestion comes from a particular source file. These files tend to be specific to substance classifications - foods, drug ingredients, drug classes, products, e.g., so the source file could in principle be deduced from the coded value. Since this ability to deduce is not a requirement, and it has not been tested, this field captures which file was used to provide the selected code.